Writing Wednesday: More on the Levigation Pool

I am committed to finishing the article that I started this winter by the end of the upcoming weekend so we can move on to more interesting (or at least less familiar) material at Polis.

My current writing has focused on clarifying the structure and chronology of what we have identified as a levigation pool near a pottery kiln at Polis. Some of what I’ve included below is similar to what I wrote earlier this year, but I managed to develop some of the description more fully (as well as the analysis of finds; you’ll notice that some of the finds information is not quite complete or legible here). 

We’ve also started to play with using ChatGPT to clean up sketches in the notebooks and even export these sketches as vector files. This is particularly important for us as we study features like the levigation pool which never received formal documentation by the project’s architects. The images below were originally on densely gridded notebook paper. ChatGPT was able to extract the line drawing from the notebook page and supply a slightly suspect scale to it. I was also able to ask ChatGPT to eliminate all text from these images and export then as SVG files for further refinement in Adobe Illustrator. Here, I could superimpose the vector image atop the original notebook illustration and fix misinterpretations by the robot. 

The Levigation Pool

Excavations in both 1990 and 1991 revealed the top of a series of eleven vertically arranged terracotta tiles arranged to form the coping of a pool. The excavators described this pool as a “fish pond,” but its proximity to a workshop area, size, and general characteristics make it more likely to have been a pool for the mixing of raw clay and water as part of the levigation process. The levigation pool itself consists of two perpendicular lines of coping tiles suggesting a pool that measured 5 meters east-west and perhaps 3 meters north-south. The tiles that defined the pool appear to be in reuse and from monumental buildings. The largest preserved tiles measured over a half a meter per side (e.g. AT54 and AT53) are 5 cm thick. Two more fragmentary tiles (AT60 and AT55) are damaged examples of this same type. These heavy, thick tiles were probably paving tiles. The rest of the larger fragments used to line the pool are roof tiles probably dating to the Hellenistic or earlier Roman period. Their 2 cm thickness with 4 cm ridges aligns well with the contemporary tiles at Panayia-Ematousa (Rose 2006). It seems likely that these tiles came from a monumental Hellenistic building at Polis and perhaps even the nearby structure at E.G0 (Najbierg 2012). The presence of hypocaust tile fragment (AT58) indicates that by the first or early second century a Roman style bath existed at Polis.

The tiles were set into a vertical bedding of clay which would have provided a degree of waterproofness. A layer of bright red clay covered the inner surface of the tiles. The interior layer of clay was either residue from the levigation process or, as plausibly, applied as a way to waterproof the sides of the pool. The floor of the pool seems to consist of a “hard, crusty, dried, clay-like earth floor” which perhaps represents the residue of the levigation process. The lack of complete waterproofing of the pool may have facilitated the draining of water and been an asset rather than a liability. Finally, there appears to be a gap in the tile wall of the basin on its southern side where the hypocaust (AT58, AT68, and AT67). This gap would have presumably been closed, perhaps with a tile, when the basin was in use and perhaps facilitated drainage after levigation occurred.

This estimate of the pool’s north-south dimensions depends upon the location of a basin described in the notebook as a “pithos.” It appears, however, that this “pithos” was an open ceramic basin whose lowest level is at approximately the same elevation as the lowest parts of the pool’s coping tiles (around 18 m ASL). It seems reasonable to conclude that the basin and the pool are contemporary and that the pool’s northern limits must fall before the area around the basin. Published examples of levigation pools from the Levant suggest that a pool of 15 sq. m would be relatively large, but not outside the range of dimensions for these features. Its size indicates the capacity to levigate large quantities of clay at the preliminary stage ceramic production process, and this suggests that it served a large production site. The basins to the north of the pool is a common feature at ceramic production sites in the Levant and should probably be associated with the treading of clay.

The main challenge in dating this pool and its associated basin is that the area where it stood was a busy one in the Roman and Late Roman periods. Not only was the basilica leveling fill and contemporary foundations cut through the area most likely disturbing the western side of the levigation pool, but a wall preserved in at least two and possible three phases seems to have complicated and potentially disturbed the southern and eastern side of the pool prior to the basilica construction. As a result, the overall stratigraphy of the area is compromised with only small areas of undisturbed soil behind the coping tiles of the pool. This area preserved a small assemblage of datable material (S06.1991.L21). The latest sherd in this material is a Cypriot Sigillata form 30 which is in the “late series” of CS forms largely dating to the first half of the 2nd century or later (Cat. 36 and 37; Hayes 1991, 44). There are also the thickened rims of contemporary globular cooking pots known from Paphos in this assemblage (Cat. 71). Eastern Sigillata Form 4 (Cat. 10) and 22 (Cat. 13) appear to date to a century earlier as does a lagynos with a parallel at Paphos (Cat. 82; Hayes Series 5, no. 14). Material excavated from within the basin (S06.1991.L17) offers little additional evidence for date although the utility and cooking wares and color coated wares might date earlier than the material associated with the pool. Since the pool and the basin are almost certainly contemporary, the presence of earlier material in the basin itself reflects the generally confused stratigraphy of the area.

The pool stood near the intersection of two walls: one to the south and one to the east. The date and precise relationship between these walls and pool remains unclear. The wall immediately to the south of the levigation pool has three phases with the later two phases being seemingly Late Roman in date, but presumably before the leveling and filling of the area for the construction of the basilica. Unfortunately, the exact relationship between the wall and the pool remains opaque as does the relationship between the south wall and east wall (and the east wall and the basin). That said, the gap in the tiles along the southern wall of the levigation pool might suggest that the south wall post-dated the pool; if the gap in the tiles was left to drain the basin when levigation was complete. Moreover, it seems unlikely that the pool was inside an enclosed space. This suggests that the south and east walls post date the pool and basin. Whatever the case, the material associated with the south wall is not substantially later than that found behind the tile packing of the pool suggesting that whatever their sequence, their chronological dating was likely quite close.

The removal of the south wall (S06.South Wall) contains CS12 (Cat. 22, Cat 24, and Cat. 25) and CS29 (Cat. 33 and Cat. 35) which generally belong to the later sequence of this type of pottery as well as CS18 (Cat. 26), pinch handled amphora (Cat. 55) and various medium coarse (Cat. 79 and 80 and kitchen wares (Cat. 66). This wall cuts through a series of levels — most of which are only narrow lenses of soil, cuts, and floors — that produced diagnostic cooking wares datable to the 1st and 2nd centuries (S06.L28P1B24 [Hayes, Paphos III, fig. 33:5]; S06.L28P1B25[Rowe 67.1]); S06.L31P2B21;S06L25P1B28 and 29). These levels also contained two fragments of potstands (S06.L25 and Cat. 84). The wall that runs along the eastern side of the levigation pool is a series of superimposed walls similar to the south wall. The latest phase of this wall is Late Roman, but the earlier phase is likely contemporary with the Roman period road to the east and perhaps joins with the south wall (see below). Our inability to distinguish the exact sequence of construction reflects the sometimes rapid adaptation of this area for new uses as production needs required.